Electric Sheep Alternative Credit Dataset v1.0

Four regions: US, NG, IN, BR
19 behavioral features for credit scoring
Two labeling strategies (rule-based + simulated)
120 users, 105K+ transactions

.gitattributes

@@ -0,0 +1,10 @@
+# Git LFS for large files
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.hdf5 filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.csv filter=lfs diff=lfs merge=lfs -text

BR/features.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84db2939337bf8114568d07a21dacd421540447b245cce4826dc751a8821507b
+size 17597

BR/labels_v1.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b2676cf3527726ae0632f00a08ba408a20832c596bbd0ba8037d7525753434e
+size 3743

BR/labels_v2.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:dc0bfd085d3c523f896e7f9325887204b9e37fd15964c6a8407c8d27fe86f0ef
+size 4378

BR/metadata.json

@@ -0,0 +1,25 @@
+{
+  "region": "BR",
+  "currency": "BRL",
+  "generated_at": "2026-03-21T13:48:45.198247",
+  "n_users": 30,
+  "n_transactions": 26350,
+  "seed": 42,
+  "label_method": "both",
+  "generation_time_seconds": 5.28,
+  "user_metadata": {
+    "region": "BR",
+    "currency": "BRL",
+    "n_users": 30,
+    "seed": 42,
+    "archetype_distribution": {
+      "stable_salaried": 8,
+      "financial_stress": 6,
+      "student": 5,
+      "gig_worker": 5,
+      "business_owner": 3,
+      "cash_heavy": 2,
+      "professional": 1
+    }
+  }
+}

BR/transactions.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0a3970379933a17c082b3ef0e80133c35197a9aca3b83b2c8b6120fb2b720143
+size 1537028

BR/users.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:300f486ee6d8391f5b01c7b6870c278f28869267c643b670a7b61ca75b2fdc51
+size 5618

IN/features.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12f4a6f23efc47c2bdc1a3f0fd88c11da469fc7c038f94fa1de8dc6334e96eb4
+size 17506

IN/labels_v1.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c85752f287850530f32bc6da88f515f0e37b3de4dd5f7a8a9096679c273ac8f1
+size 3759

IN/labels_v2.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8f54bdba67e6d839069fb778eb7a62535ee98e722eff204e189cd91851ada086
+size 4391

IN/metadata.json

@@ -0,0 +1,25 @@
+{
+  "region": "IN",
+  "currency": "INR",
+  "generated_at": "2026-03-21T13:48:39.915667",
+  "n_users": 30,
+  "n_transactions": 25526,
+  "seed": 42,
+  "label_method": "both",
+  "generation_time_seconds": 5.07,
+  "user_metadata": {
+    "region": "IN",
+    "currency": "INR",
+    "n_users": 30,
+    "seed": 42,
+    "archetype_distribution": {
+      "stable_salaried": 8,
+      "cash_heavy": 6,
+      "student": 5,
+      "gig_worker": 5,
+      "business_owner": 3,
+      "financial_stress": 2,
+      "professional": 1
+    }
+  }
+}

IN/transactions.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a2ce89a11182ff1307ec648efbae80dc79f9012adbb9c9372994214cfc5ebf5c
+size 1476997

IN/users.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9b181b94396698c35a6adc0860624d9d88c079919b1b84a960144c9bc0d9276
+size 5617

LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Electric Sheep Contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

NG/features.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c417fe144f665ed1027101b0032e9a09a20bc1e884a244d72bb55a0fd85c913b
+size 17557

NG/labels_v1.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5e5e9be49c6009ea96dc27aab5b070016dbd2da5e15a2b96717f0662912a3ad
+size 3761

NG/labels_v2.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fdad453ada74446b405a6e7fede2c817e6cbbe7ee3de7c4537f3159807b29c5a
+size 4385

NG/metadata.json

@@ -0,0 +1,25 @@
+{
+  "region": "NG",
+  "currency": "NGN",
+  "generated_at": "2026-03-21T13:48:34.845351",
+  "n_users": 30,
+  "n_transactions": 25402,
+  "seed": 42,
+  "label_method": "both",
+  "generation_time_seconds": 5.18,
+  "user_metadata": {
+    "region": "NG",
+    "currency": "NGN",
+    "n_users": 30,
+    "seed": 42,
+    "archetype_distribution": {
+      "stable_salaried": 8,
+      "cash_heavy": 6,
+      "student": 6,
+      "gig_worker": 4,
+      "business_owner": 3,
+      "financial_stress": 2,
+      "professional": 1
+    }
+  }
+}

NG/transactions.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bde733731a829c435c2d4f6ed17343760903a227be55e9f31d0dce62f3b658c0
+size 1474138

NG/users.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e4f95d3051224e5423dcd5d42cd1684dfc249fecb97904d7a4a4990f5590c1f
+size 5619

README.md

@@ -0,0 +1,241 @@
+---
+annotations_creators:
+- machine-generated
+language:
+- en
+license:
+- mit
+multilinguality:
+- monolingual
+pretty_name: Electric Sheep Alternative Credit Data for Thin-File Users
+size_categories:
+- 100<n<1000
+source_datasets:
+- original
+tags:
+- credit-scoring
+- financial-inclusion
+- thin-file
+- synthetic-data
+- behavioral-finance
+- alternative-data
+- tabular
+- fintech
+- machine-learning
+task_categories:
+- tabular-classification
+- tabular-regression
+configs:
+- config_name: US
+  data_files: "US/*.parquet"
+- config_name: NG
+  data_files: "NG/*.parquet"
+- config_name: IN
+  data_files: "IN/*.parquet"
+- config_name: BR
+  data_files: "BR/*.parquet"
+---
+
+# Electric Sheep — Alternative Credit Data for Thin-File Users
+
+## Dataset Summary
+
+The **Electric Sheep Alternative Credit Dataset** is a synthetic, multi-layer dataset designed to model creditworthiness using behavioral financial data rather than traditional credit history. It targets **thin-file users** — individuals excluded from conventional credit scoring systems due to insufficient loan or credit card history.
+
+This dataset provides **culturally authentic financial behavior data** for four distinct regions:
+
+| Region | Currency | Channel Dominant | Income (median) |
+|--------|----------|------------------|-----------------|
+| **US** (United States) | USD | POS / Bank | ~$17,400 |
+| **NG** (Nigeria) | NGN | Mobile Money | ~₦2,400 |
+| **IN** (India) | INR | UPI | ~₹1,900 |
+| **BR** (Brazil) | BRL | PIX / Bank | ~R$3,500 |
+
+## Supported Tasks
+
+- **Binary Classification:** Creditworthy vs not creditworthy
+- **Multi-class Classification:** Good / Bad / Indeterminate credit outcome
+- **Regression:** Default probability estimation
+- **Fairness Research:** Bias analysis across demographic groups
+- **Cross-Regional Studies:** Behavioral pattern comparison across economies
+
+## Dataset Structure
+
+Each region contains **5 parquet files**:
+
+| File | Rows | Description |
+|------|------|-------------|
+| `users.parquet` | 1 per user | Demographic metadata (age, income type, archetype) |
+| `transactions.parquet` | ~900 per user | Raw financial behavior (amounts, categories, channels) |
+| `features.parquet` | 1 per user | 19 aggregated behavioral metrics |
+| `labels_v1.parquet` | 1 per user | Rule-based credit labels |
+| `labels_v2.parquet` | 1 per user | Simulated loan outcome labels |
+
+### Users Table
+
+| Field | Type | Description |
+|-------|------|-------------|
+| user_id | string | Unique identifier (format: `{region}_{index:06d}`) |
+| age_range | string | Age bucket: `18-25`, `26-35`, `36-45`, `46-55`, `55+` |
+| income_type | string | `salary`, `gig`, `business`, `unemployed` |
+| region | string | `US`, `NG`, `IN`, `BR` |
+| currency | string | `USD`, `NGN`, `INR`, `BRL` |
+| account_tenure_days | int | Length of financial activity in days |
+| primary_archetype | string | Dominant behavioral pattern |
+| secondary_archetype | string | Secondary behavioral overlay |
+
+### Transactions Table
+
+| Field | Type | Description |
+|-------|------|-------------|
+| transaction_id | string | Unique transaction ID (UUID) |
+| user_id | string | Foreign key to users |
+| timestamp | datetime | Transaction time (2024-01-01 to 2025-12-31) |
+| amount | float | Signed: `+income`, `-expense` in local currency |
+| transaction_type | string | `credit` or `debit` |
+| category | string | `food`, `transport`, `bills`, `entertainment`, `betting`, `transfer`, `savings`, `healthcare`, `other` |
+| channel | string | `bank`, `cash`, `POS`, `mobile_money` |
+| merchant_type | string | Region-specific merchant classification |
+| balance_estimate | float | Running balance approximation |
+| is_recurring | bool | Recurring payment flag |
+| counterparty | string | Transfer recipient/sender (if applicable) |
+| corridor | string | Remittance corridor (e.g., `US→NG`) |
+
+### Features Table
+
+| Field | Type | Description |
+|-------|------|-------------|
+| user_id | string | Foreign key |
+| avg_monthly_income | float | Mean monthly inflow |
+| income_volatility | float | Coefficient of variation of monthly income |
+| income_frequency | float | Income events per month |
+| income_trend | float | Linear trend (positive = growing) |
+| income_gap_months | int | Months with zero income |
+| avg_monthly_spend | float | Mean monthly outflow |
+| spending_volatility | float | Variability of monthly spending |
+| essential_spend_ratio | float | Fraction on food, transport, bills, healthcare |
+| discretionary_spend_ratio | float | Fraction on non-essentials |
+| betting_ratio | float | Fraction on betting/gambling |
+| avg_balance | float | Mean balance over period |
+| min_balance | float | Lowest balance reached |
+| max_balance | float | Highest balance reached |
+| overdraft_frequency | float | Fraction of transactions where balance < 0 |
+| bill_payment_consistency | float | Regularity of bill payments (0–1, higher = better) |
+| recurring_expense_ratio | float | Fraction of spending that recurs |
+| transaction_frequency | float | Transactions per day |
+| cashflow_stability_score | float | Composite score (0–1, higher = more stable) |
+| risk_behavior_score | float | Composite risk indicator (0–1, higher = riskier) |
+
+### Labels Table
+
+| Field | Type | Description |
+|-------|------|-------------|
+| user_id | string | Foreign key |
+| credit_outcome | string | `good`, `bad`, `indeterminate` |
+| default_probability | float | Estimated probability of default (0–1) |
+| risk_bucket | string | `low`, `medium`, `high` |
+| labeling_method | string | `v1_rules` or `v2_simulation` |
+
+## Loading the Dataset
+
+### Python (Hugging Face Datasets)
+
+```python
+from datasets import load_dataset
+
+# Load US region
+us = load_dataset("electricsheepafrica/electric-sheep-credit", name="US")
+users = us["train"]["users"]
+transactions = us["train"]["transactions"]
+features = us["train"]["features"]
+labels = us["train"]["labels_v2"]
+
+# Load Nigeria
+ng = load_dataset("electricsheepafrica/electric-sheep-credit", name="NG")
+
+# Load India
+ind = load_dataset("electricsheepafrica/electric-sheep-credit", name="IN")
+
+# Load Brazil
+br = load_dataset("electricsheepafrica/electric-sheep-credit", name="BR")
+```
+
+### Python (Pandas)
+
+```python
+import pandas as pd
+
+users = pd.read_parquet("US/users.parquet")
+transactions = pd.read_parquet("US/transactions.parquet")
+features = pd.read_parquet("US/features.parquet")
+labels = pd.read_parquet("US/labels_v2.parquet")
+```
+
+## Feature Correlations with Creditworthiness
+
+Strong and significant correlations (point-biserial r, p < 0.001 unless noted):
+
+| Feature | US | NG | IN | BR | Interpretation |
+|---------|----|----|----|----|----------------|
+| `overdraft_frequency` | −0.73*** | −0.84*** | −0.80*** | −0.74*** | Strongest negative predictor |
+| `cashflow_stability_score` | +0.58*** | +0.70*** | +0.71*** | +0.66*** | Strongest positive predictor |
+| `betting_ratio` | −0.36* | −0.60*** | −0.65*** | −0.60*** | Significant negative |
+| `avg_balance` | +0.49** | +0.56** | +0.40* | +0.48** | Positive (liquidity buffer) |
+| `bill_payment_consistency` | +0.50** | +0.50** | +0.38* | +0.37* | Reliability signal |
+
+## Dataset Statistics (per region, 30 users)
+
+| Metric | US | NG | IN | BR |
+|--------|----|----|----|----|
+| Transactions per user | ~936 | ~847 | ~851 | ~878 |
+| Spend/Income ratio (median) | 0.93 | 1.06 | 0.98 | 1.01 |
+| Users ending with negative balance | 8/30 | 13/30 | 10/30 | 10/30 |
+| Users who had overdraft | 10/30 | 18/30 | 17/30 | 19/30 |
+| Peak spending month | December | December | October | February |
+| Dominant payment channel | POS/Bank | Mobile Money | UPI | PIX/Bank |
+
+## Generation Methodology
+
+This dataset is generated using a **hybrid synthetic-realistic pipeline**:
+
+1. **User Archetypes**: 8 behavioral templates (stable salaried, gig worker, business owner, student, cash-heavy, financial stress, professional, gambler) sampled with region-specific weights
+2. **Transaction Simulation**: Poisson-based timing, log-normal amounts, category-specific spending profiles
+3. **Noise Injection**: Amount jitter, category misclassification, timestamp drift, duplicate detection
+4. **Feature Engineering**: 19 behavioral metrics computed over rolling time windows
+5. **Label Assignment**: Rule-based heuristics (v1) and simulated loan outcomes (v2)
+
+## Limitations
+
+- **Synthetic data**: Not from real financial institutions
+- **Heuristic labels**: Credit outcomes are simulated, not observed
+- **Simplified patterns**: Real financial behavior has more complex temporal patterns
+- **Regional estimates**: Income levels and spending patterns are approximated
+
+## Use Cases
+
+- Training credit scoring models for thin-file populations
+- Benchmarking alternative data approaches
+- Research on financial inclusion
+- Fairness and bias testing across demographic groups
+- Cross-regional behavioral finance studies
+
+## Citation
+
+```bibtex
+@dataset{electric_sheep_2026,
+  title={Electric Sheep: Alternative Credit Data for Thin-File Users},
+  author={ElectricSheepAfrica},
+  year={2026},
+  version={1.0},
+  license={MIT},
+  url={https://huggingface.co/datasets/electricsheepafrica/electric-sheep-credit}
+}
+```
+
+## License
+
+MIT — free for research and commercial use.
+
+## Contact
+
+[ElectricSheepAfrica](https://huggingface.co/electricsheepafrica)

US/features.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c4685a4ca8e488ef30798849f2ddffd19ffe0ccd3f29124529b4d94f7d8df531
+size 17537

US/labels_v1.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:952084b91ffbd8d3beb063ca63fd80711fd6c1693c679960c4d21b53e70988d8
+size 3737

US/labels_v2.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4689e7259f43b467c065c9cb2efee57b06d5c9369b713bd1061a585ffd31d0bf
+size 4395

US/metadata.json

@@ -0,0 +1,24 @@
+{
+  "region": "US",
+  "currency": "USD",
+  "generated_at": "2026-03-21T13:48:29.655879",
+  "n_users": 30,
+  "n_transactions": 28084,
+  "seed": 42,
+  "label_method": "both",
+  "generation_time_seconds": 5.67,
+  "user_metadata": {
+    "region": "US",
+    "currency": "USD",
+    "n_users": 30,
+    "seed": 42,
+    "archetype_distribution": {
+      "stable_salaried": 9,
+      "financial_stress": 6,
+      "gig_worker": 6,
+      "student": 5,
+      "professional": 3,
+      "business_owner": 1
+    }
+  }
+}

US/transactions.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4f1eb2657dbce6a13df1f2301c1daec19fe33aed9d612be456497dec8bcbfac
+size 1684389

US/users.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ec28c22b2b5b076f86a8ab4302241edbc7029aaca9015ba2afdf51ae59dcceeb
+size 5607